Towards Dynamic Load Balancing in Heterogeneous Cluster Using Mobile Agent

Dynamic load balancing (DLB) plays a critical role in developing high performance cluster. It is an important system function to distribute workload among available processors to improve throughput. Existing load balancing techniques are mainly concerned about the use of two main resources namely CPU and Memory. But there are certain I/O data Intensive applications where load balancing becomes important issue. Due to the ignorance of I/O resources, the previous CPU or memory based load balancing techniques suffer a major performance drop. Hence in this paper we propose Mobile Agent Based Load Balancing (MALB) algorithm for these types of applications. In order to distribute the incoming load on participating nodes, we use mobile agent (MA). A value function which is sum of all three types of resources including CPU, Memory and I/O is considered. The MA choose this value function as a trademark for load transfer. Different metrics are used to compare load balancing mechanism with the existing technology. The experiment is carried out on cluster of PC´s divided into multiple LAN´s using PMADE (Platform for Mobile agent distribution and execution).